Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add ensemblr-hq/ensemblr --skill architecture-diagramgit clone --depth 1 https://github.com/ensemblr-hq/ensemblrWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ensemblr-hq/ensemblr/architecture-diagram)<a href="https://agentmods.dev/skills/ensemblr-hq/ensemblr/architecture-diagram"><img src="https://agentmods.dev/badge/skills/ensemblr-hq/ensemblr/architecture-diagram/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ensemblr-hq/ensemblr/architecture-diagram"><img src="https://agentmods.dev/badge/skills/ensemblr-hq/ensemblr/architecture-diagram.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00094 | $0.02769 |
| Opus 5 | $0.00047 | $0.01385 |
| Sonnet 5 | $0.00019 | $0.00554 |
| Haiku 4.5 | $0.00009 | $0.00277 |
Grade A, and why
architecture-diagram scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 238 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Drawing the architecture diagram
Ensemblr keeps an architecture diagram of every workspace at
.ensemblr/architecture.json, committed alongside the code it describes. The
user opens it from the tab strip.
Nothing derives it. The app ships no scanner: a workspace nobody has drawn has no diagram at all, and its pane shows an empty state whose button opens a fresh chat and asks for one — which may well be how you got here. Producing it is entirely your work, and what you store is what the user sees until somebody stores something else.
Read and write it through the two ops below rather than by editing the file directly — they validate the document, and the update op refreshes any diagram tab the user has open.
So the job is one of two, depending on what the read answers: draw one from the codebase when there is none, or edit the one that exists when there is. Either way the target is a drawing a maintainer would make, not a dump of the folder tree.
It is a drawing for people, not a map for you
The diagram exists so a human can see the shape of the repository at a glance. It is not a source of truth about the code, and you must never treat it as one. You curate it; you do not consult it.
Concretely: never answer a question about the codebase from the diagram, never decide what to edit or where something lives because a node says so, and never report its contents as fact. It is lossy by design — nodes are dropped as noise, edges are omitted, labels are renamed to concepts that appear nowhere in the source — and it is only as current as the last agent who bothered to update it, which may be nobody since the code moved.
The codebase is the truth. Read the files. If the diagram and the source disagree, the diagram is wrong, and fixing it is the only thing that disagreement licenses you to do.
The two calls
ensemblr_get_architecture_diagram({}) → the document to edit
ensemblr_update_architecture_diagram({ diagram }) → store your edit
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 238 lines · 94 tokens per session scan A d26c42c2c2e7
architecture-diagram is a skill published in the GitHub repository ensemblr-hq/ensemblr (8 stars, last pushed today), licensed Apache-2.0. It adds 94 tokens to every session and 2,769 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-06.
Other skills, from other repositories
frame-macos-notification
Realistic macOS notification banner with app icon, title, and body, suited to video overlays or product teasers.
hig-review
Reviews and improves any UI against Apple's Human Interface Guidelines — accessibility, platform conventions, layout, color and type, materials and Liquid Glass, interaction states, and writing — citing the specific guideline behind every finding and expressing every fix in the user's own stack (web, Electron, React…
design-qa
Compares rendered prototype screenshots side-by-side against the pinned target design, iteratively eliminating visual discrepancies across 5 fidelity surfaces.
ideate
Generates exactly three distinct visual design options with independent layouts and interaction models, iterates on user feedback, and pins the confirmed option as the visual target.
audit
Captures and reviews product journeys, funnels, screens, and multi-step user experiences based on screenshot evidence, evaluating UX, visual design, and accessibility.
motion
Injects high-craft UI motion, spring physics, tactile micro-interactions, page transitions, and staggered animations into components and prototypes.